A compiler, in general, reads higher level language computer code and converts it to either p-code or native machine code. An interpreter runs directly from p-code or an interpreted code such as Basic or Lisp. Typically, compiled code runs much faster, is more compact, and has already found all of the syntax errors and many of the illegal reference errors. Interpreted code only finds such errors after the application attempts to interpret the affected code. Interpreted code is often good for simple applications that will only be used once or at most a couple times, or maybe even for prototyping. Compiled code is better for serious applications. A compiler first takes in the entire program, checks for errors, compiles it and then executes it. Whereas, an interpreter does this line by line, so it takes one line, checks it for errors, and then executes it.
If you need more information, just Google for "difference between compiler and interpreter".